Nvidia refutes report of halting AI compute partnership deals

Nvidia denied a Wall Street Journal report that it paused some transactions under its AI Compute Partnership program. The company said the initiative, introduced in July, remains active and is evolving due to high demand. The report claimed partners were unhappy with Nvidia's attempts to control GPU leasing and raised antitrust concerns.

EXPANDED:

The AI Compute Partnership program, introduced in July, involves Nvidia committing to rent capacity at newly built AI data centers and guaranteeing minimum revenue levels. This structure helps operators secure financing for billion-dollar facilities before customer contracts are in place, essentially using Nvidia's demand commitments as collateral for lenders.

The reported friction centers on Nvidia's attempts to dictate which customers could lease its GPUs, preferring to distribute capacity across multiple smaller AI firms rather than concentrating it with single large clients. Cloud providers pushed back, asserting their right to choose their own customers. Nvidia's denial suggests the program's terms are being adjusted rather than abandoned, with the company citing high demand as the reason for evolution